Personalized Chemotherapy Profiling Using Cancer Cell Lines from Selectable Mice

Published OnlineFirst January 22, 2013; DOI: 10.1158/1078-0432.CCR-12-2127

Clinical Cancer Cancer Therapy: Preclinical Research

Personalized Chemotherapy Proﬁling Using Cancer Cell Lines from Selectable Mice

Hirohiko Kamiyama1, Sherri Rauenzahn1, Joong Sup Shim2, Collins A. Karikari1, Georg Feldmann1, Li Hua1, Mihoko Kamiyama1, F. William Schuler3, Ming-Tseh Lin1, Robert M. Beaty1, Balasubramanyam Karanam1, Hong Liang1, Michael E. Mullendore1, Guanglan Mo3, Manuel Hidalgo3, Elizabeth Jaffee3, Ralph H. Hruban1,3, H.A. Jinnah4, Richard B.S. Roden1, Antonio Jimeno3, Jun O. Liu2, Anirban Maitra1,3, and James R. Eshleman1,3

Abstract Purpose: High-throughput chemosensitivity testing of low-passage cancer cell lines can be used to prioritize agents for personalized chemotherapy. However, generating cell lines from primary cancers is difficult because contaminating stromal cells overgrow the malignant cells. Experimental Design: We produced a series of hypoxanthine phosphoribosyl transferase (hprt)-null immunodeficient mice. During growth of human cancers in these mice, hprt-null murine stromal cells replace their human counterparts. Results: Pancreatic and ovarian cancers explanted from these mice were grown in selection media to produce pure human cancer cell lines. We screened one cell line with a 3,131-drug panel and identified 77 U.S. Food and Drug Administration (FDA)–approved drugs with activity, and two novel drugs to which the cell line was uniquely sensitive. Xenografts of this carcinoma were selectively responsive to both drugs. Conclusion: Chemotherapy can be personalized using patient-specific cell lines derived in biochemically selectable mice. Clin Cancer Res; 19(5); 1139–46. 2012 AACR.

Introduction to respond to specific drugs. In vitro chemotherapy sensi- Selecting the most appropriate chemotherapy for a given tivity and resistance assays (CSRA), a natural extrapolation patient has historically been based on histopathology and of antimicrobial susceptibility testing, was first reported in past studies showing that specific drugs are generally active 1957 (1). Some 50 years later, however, the official position against that malignancy. However, drugs deemed active of the American Society of Clinical Oncology has been that against a particular type of cancer are most commonly this type of testing is still not ready for routine clinical use active in only a subset of patients with those cancers. (2). Problems with CSRAs include the many different for- Furthermore, chemotherapeutic agents are generally more mats for this testing (2), and results from clinical trials, toxic and expensive than most other medications in wide- comparing CSRA-guided therapy with conventional treat- spread use. Accordingly, an alternative strategy is to test ment, have varied from no significant difference to highly functionally whether a given patient’s cancer is likely or not significant benefit (3). Finally, most groups test cancer cells immediately after resection, and these may be sick or dying from hypoxia, anesthetic drugs, or overnight shipping and Authors' Affiliations: The Sol Goldman Pancreatic Cancer Research Center, 1Departments of Pathology, 2Pharmacology, and 3Oncology, so any toxicity to these cells may reflect synergistic toxicity of Johns Hopkins University School of Medicine, Baltimore, Maryland; and the drugs tested with any of these effects. 4 Department of Neurology, Emory University School of Medicine, Atlanta, We hypothesize that low-passage cell lines might better Georgia represent their respective tumors and therefore more accu- Note: Supplementary data for this article are available at Clinical Cancer rately predict in vivo chemosensitivity. Isolating cancer cell Research Online (http://clincancerres.aacrjournals.org/). lines, however, can, counterintuitively, be difficult, espe- H. Kamiyama and S. Rauenzahn contributed equally to this work. cially from solid primary tumors (4). To date, the success rate for the generation of cell lines is only 10% to 40% for A. Maitra and J.R. Eshleman contributed equally to this work. many solid tumors (4–8). The most significant barrier to The hprt-null NSG mice are available from Jackson Labs (Bar Harbor, routine cancer cell line production is that when tumors are Maine) as mouse #012480 (NOD.Cg-Prkdc IL2Rg Hprt/EshJ). explanted into tissue culture, fibroblasts and other stromal cells proliferate, overgrow, and eliminate the malignant Corresponding Author: James R. Eshleman, The Sol Goldman Pancreatic Cancer Research Center, CRB II, Suite 344, Johns Hopkins University cells. School of Medicine, 1550 Orleans Street, Baltimore, MD 21231. Phone: We report production of nude-, severe combined immu- 410-955-3511; Fax: 410-614-0671; E-mail: [email protected] nodeficiency- (SCID), and NSG- [Nonobese diabetic doi: 10.1158/1078-0432.CCR-12-2127 (NOD), SCID, interleukin-2 receptor gamma (IL2Rg)], hprt- 2012 American Association for Cancer Research. null mice. During growth of implanted human cancers in

www.aacrjournals.org 1139

Kamiyama et al.

flow biosafety cabinet, tumors were finely minced and Translational Relevance incubated in Dulbeccos’ Modified Eagles’ Media (DMEM) Cancer cell lines are essential for functional studies of containing 750 U/mL of type IV collagenase (Invitrogen) cancer. When explanted tumors are grown in vitro, how- and 500 U/mL of hyaluronidase (Sigma-Aldrich Inc.) for 1 ever, noncancer fibroblasts paradoxically overgrow the hour at 37C and pipetting up and down 20 times to malignant cells. Here, we report production of a series of mechanically fragment the tumor. The cell solution was immunodeficient hypoxanthine phosphoribosyl transferase plated at 0.5, 1, and 2 mL per well on 6-well plates coated (hprt)-null mice. When human tumors are expanded in with or without type I rat tail collagen (BD Biosciences) in these mice, the resultant tumor is composed of human minimum essential media (MEM) with 20% FBS, supple- cancer cells and biochemically defective mouse stromal mented with standard concentrations of penicillin/strepto- cells. When these tumors are explanted into culture mycin, L-glutamine, 5 ng/mL of human recombinant EGF and grown in hypoxanthine–aminopterin–thymidine (all Invitrogen), and 0.2 U/mL of human insulin (Sigma- (HAT) media, human cancer cell lines can be readily Aldrich). Following 2 days of recovery (designated day 0), isolated. We show proof-of-principle of the mice for 1 HAT (100 mmol/L sodium hypoxanthine, 400 nmol/L personalized cell line isolation and chemosensitivity aminopterin, 16 nmol/L thymidine; Invitrogen) media testing. This approach may be used to guide chemother- were fed to half of the cultures. Cells are fed with or without apy selection in the future. HAT media every 2 to 3 days. All cell lines were confirmed as patient in origin by DNA fingerprinting with the Power- plex 1.2 Kit per manufacturers’ instructions (Promega). They were also stained with anti-cytokeratin (Ventana, these mice, biochemically defective mouse stromal cells 760-2595) and anti-smooth muscle actin (Ventana, 760- replace the human stromal cells from the original tumor 2835) by immunohistochemistry and shown to be tumor- (9, 10). After explanting these xenografts into culture, mouse igenic in nude mice. cells can be eliminated by selection in hypoxanthine, aminopterin, and thymidine (HAT) containing media (11, 12) Chemosensitivity testing—initial library screening and to isolate pure human cancer cells. Using this system, we confirmation isolated a total of 6 pancreatic ductal adenocarcinoma Cancer cells were screened for chemosensitivity using the (PDA) cell lines and 1 ovarian cancer cell line. One of the Johns Hopkins Drug Library (JHDL) consisting of 3,131 PDA cell lines was isolated from the corresponding surgically drugs arrayed in 96-well plates. This includes 1,907 U.S. resected sample and tested for in vitro chemosensitivity using Food and Drug Administration (FDA)-approved drugs, 570 a 3,131-drug panel (13). In comparison to other PDA cell drugs approved in other countries, and 654 other drugs in lines, this cell line in vitro was differentially more sensitive to various phases of development (13). All drugs are main- digitoxin and nogalamycin, which correlated with in vivo tained at 70 C at a stock concentration of 200 mmol/L in response in mice, where the same 2 drugs showed selective PBS containing 2% dimethyl sulfoxide (DMSO) and 10% activity against xenografted tumors. These data suggest a FBS. possible novel paradigm for functional personalized che- Tumor cells were subcultured, counted, and inoculated motherapeutic selection by isolating low-passage cancer cell into 96-well plates at 5,000 cells per well in complete media lines and screening them with large drug panels. and allowed to adhere overnight. Drugs were then added to the plates in duplicate wells at a final concentration of 10 m Materials and Methods mol/L for a total of 48 hours. After the first 24 hours in drug, 1 mCi [3H]-thymidine was added and incubated for an Patient samples and xenografting additional 24 hours (in both drug and [3H]-thymidine). Primary tumors from patient-derived resection speci- Cells were then trypsinized using 1 trypsin-EDTA (Invi- mens or xenografts from standard nude mice were harvested trogen). The suspended cells were transferred onto Filter- with informed consent and Institutional Review Board Mat-A glass fiber filters (Wallac) using the Harvester-96 cell (IRB) approval. They were implanted in anesthetized stan- harvester (Tomtec). The glass fiber filters were washed 5 dard B6 nude and hprt-null B6 nude mice. Following a small times with water to wash away unincorporated free [3H]- 3 skin incision, a 5-mm sample coated with Matrigel Matrix thymidine. On each plate, the first and last columns were (BD Biosciences) was implanted subcutaneously in each used as no-drug controls. Incorporated thymidine was read upper flank, the incision was closed using staples, and the on a MicroBeta plate reader (Perkin Elmer), and the mean site was disinfected with Betadine (Purdue Frederick). percentage of inhibition was calculated relative to no-drug Among the 6 pancreatic cancers, 3 were selected because control wells. A histogram of the percentage of inhibition they were familial and 3 because they have undergone full for the drugs in the JHDL was established using 2.5% size exomic DNA sequencing as xenografted samples (14). bins, and normal distribution curve fitting was conducted by the least-squares method using GraphPad ver. 5.02 Tumor harvest and culture (GraphPad Software Inc.). The mean 3 SD encompasses When tumors reached approximately 1 cm in diameter, 3,014 drugs. Using Gaussian statistics and 3 SD as the cutoff mice were sacrificed and tumors harvested. In a laminar (40.4% inhibition), then we expect about 4 drugs to be

1140 Clin Cancer Res; 19(5) March 1, 2013 Clinical Cancer Research

Cancer Cell Line Isolating Tools

false-positives among the drugs showing 40% to 100% inhibition (2,817 0.0013 ¼ 3.7). A I II III IV IC50 values of selected drugs were determined from percentage of growth inhibition of 10 different concentrations of the drugs obtained in the same experimental design described above. IC50 values of each drug for each cell line were calculated using 4-parameter logistic equation with 3 replicates results by GraphPad ver. 5.02.

Growth inhibition of subcutaneous xenografted Male Female 1Kb I II III IV WT HET dH O 1 kb tumors B Null Null 2 Tumor xenografts were generated by subcutaneous injec- + Δ + Δ + Δ + Δ + Δ + Δ + Δ + Δ + Δ 6 1,000 tion of 10 10 cells bilaterally into 6-week athymic nude- 500 Foxn1nu mice (Harlan). Tumor volumes were measured, and when the tumors reached approximately 100 mm3 in size, C II (Δ/Δ) III (wt/Δ) the mice were stratified (day 0) to treatment and nontreat- ment groups with 5 mice per group so that each group was HAT equivalent based on tumor volume. Tumor volume is obtained by the formula: length (width)2 0.5, where length is the longest diameter and width is the shortest diameter perpendicular to the length. The mice received 6TG treatment with daily intraperitoneal injection of nogalamycin (0.2 and 1 mg/kg in vehicle), digitoxin (0.4 and 2 mg/kg in vehicle), or vehicle (0.9% NaCl with 1% DMSO) control for 30 days (days 1–30). Tumor volumes were measured nu nu twice a week. At the completion of the study, mice were Figure 1. Athymic hprt-null mice. A, third cross Foxn1 / mice. The nude euthanized and tumors were measured, harvested, and hprt-null mouse (II) and heterozygous littermate controls (I, III, IV) show similar phenotypic characteristics in size, hair growth cycles, and skin weighed. The tumor volume index (TVI) was determined pigmentation. B, hprt genotyping: hprt PCR products are separated on a from a ratio of the tumor volume on a given day divided by 1% agarose gel. Wild-type (þ) and null (D) alleles are assayed in separate the tumor volume of day 0. The harvested tumors were then PCR reactions. Wild-type (WT) mice result in a 746-bp band in the wild- weighed and means and standard SDs calculated. The nor- type reaction and only the control 500-bp band in the D reaction. malized tumor weight of treatment group was calculated by Homozygous null females (II, female null) and hemizygous null males (male null) result in a 628-bp band in the D reaction but no band in the wild- dividing the treatment values by the control group for each type reaction. Heterozygous females (I, III, IV, HET) result in a 746-bp cell line (i.e., the mean tumor weight of control group for band in the wild-type reaction and a 628-bp band in the D reaction. Note, a each cell line is 100). Statistical analysis was conducted using 500-bp control band is in the D reaction in all lanes. C, tail cell the unpaired Student t test on Graph Pad Prism ver. 5.02. phenotyping. Homozygous hprt-null (II) and heterozygous (III) tail snips were harvested from female mice, trypsinized, and plated in the presence of HAT- or 6TG-containing media. Bright field microscopy of crystal Results violet–stained cells. Production of hprt-null immunodeficient mice A series of hprt-null immunodeficient mice, including Patient-specific cancer cell line production nude hprt-null (Supplementary Fig. S1), SCID hprt-null A PDA xenograft (Panc410) was implanted into these (Supplementary Fig. S2), and NSG hprt-null, were produced mice and the tumors were harvested for cell culture. After by breeding immunocompetent hprt-null mice with the plating and 2 days of recovery, cultures of Panc410 (day 0) appropriate immunodeficient mice (see Supplementary showed a mixture of cancer cells clustering together to form Information). Mice that are homozygous or hemizygous "islands" and scattered fibroblasts and other stromal cells for the hprt-null allele and homozygous nu/nu appear to (Fig. 2A and D). Cultures explanted from the hprt-null have no phenotypic variation from standard Foxn1nu/nu mouse and treated with HAT selection media typically mice (Fig. 1A). To confirm genotyping (Fig. 1B, Supple- yielded fibroblast-free cancer cell lines after 2 to 3 weeks mentary Fig. S3), we harvested tail cuttings from mice II and (Fig. 2A–C). During this time, HAT selection eliminated the III (Fig. 1A), trypsinized them, and grew the cells in tissue fibroblasts and other stromal cells, allowing for the small culture. These samples were grown in the presence of HAT or cancer cell "islands" to expand and eventually coalesce. In 6-thioguanine (6TG) to select for or against HPRT function, contrast, a sister culture, grown in the absence of HAT, was respectively (Fig. 1C). As predicted from their genotype heavily contaminated with fibroblasts (Fig. 2D–F). These (Fig. 1B), cells from the hprt-null mouse died in HAT media stromal cells surrounded cancer cell islands (Fig. 2E), over- and survived in 6TG containing media. The HPRT-profi- grew them (Fig. 2F), and precluded producing cancer cell cient heterozygous hprt-null/wt mouse cells gave the oppo- lines from these cultures. The resultant cell line matched the site result, as expected. patient’s microsatelllites and was tumorigenic in nude mice.

www.aacrjournals.org Clin Cancer Res; 19(5) March 1, 2013 1141

Kamiyama et al.

Day 0 Day 10 Day 19 A B C Figure 2. Rapid pancreatic cancer cell line production. Samples HAT explanted from Panc410 tumors grown in a nude hprt-null mouse (A–F) in the presence (A–C) or absence (D–F) of HAT media and photographed using phase microscopy at T ¼ 0, 10, and 19 D E F days in culture. Note that after 19 days in the absence of HAT media No (F), ﬁbroblasts have overgrown the culture. HAT

To confirm its epithelial origin, the cell line was stained for mitoxantrone) and 2 terpenoid natural products (pristi- cytokeratin and smooth muscle actin, which stained pos- merin and triptonide). itive and negative, respectively, as expected (data not shown). Novel drugs with activity against PDA From the drugs showing 75% or higher inhibition at 10 Chemosensitivity-initial screen mmol/L final concentration, we selected 10 drugs to study One potential application of rapid cell line isolation is further (Table 1), after excluding drugs with well-documented personalized chemosensitivity testing. To show use of the activity against pancreatic cancer, and those with relatively mice for this purpose, we first implanted SCID hprt-null high toxicities (i.e., low LD50 values in mice). Only one of mice with a surgically resected PDA and isolated the cell line these drugs is FDA-approved. We generated dose–response Panc502 (Supplementary Fig. S4). We then screened this curves to confirm activity and establish IC50 values for cell line against the JHDL (13) consisting of 3,131 drugs Panc502, Panc410, and another low-passage PDA cell line, (Supplementary Fig. S5). The library includes 1,907 U.S. Panc486, described below. From these, we selected 2 drugs FDA-approved drugs, 570 drugs approved for clinical use in because they showed significant variability in their IC50s other countries, and 654 other drugs in various phases of among the 3 low¼passage PDA cell lines (Fig. 3B). The development. A histogram of the number of drugs, graphed anthracycline, nogalamycin, was approximately 11 more as a function of the percentage of growth inhibition (Fig. effective against Panc502 than either Panc410 or the non- 3A), showed that most drugs fit under a Gaussian distribu- transformed pancreatic duct cell line, HPDE (data not shown). tion (mean, 5.0%; SD, 11.8%), which we interpreted as Finally, the cardiac glycoside digitoxin was substantially more inactive drugs for this cancer cell line. The mean 3SD toxic to Panc502 than Panc410, whereas Panc486 had an encompasses 2,817 drugs or approximately 90% of those intermediate sensitivity. We also generated dose–response tested. More importantly, 314 drugs (10%) showed activ- curves using a panel of topoisomerase inhibitors, showing ity above the normal distribution, and these were consid- that topoisomerase I inhibitors were generally more effective ered potentially active drugs (Supplementary Table S1). than topoisomerase II inhibitors (Supplementary Fig. S6). These 314 drugs included 77 drugs that are currently FDA-approved and 12 drugs previously approved but cur- Does in vitro chemosensitivity predict in vivo response? rently discontinued (combined 2.8% of drugs in the To address the hypothesis that in vitro response could library). predict in vivo response, we then raised xenografted tumors The FDA-approved drug, digoxin, showed 95% growth from Panc410 and Panc502 cell lines and treated the mice inhibition, consistent with its recent discovery as a potential harboring these xenografts with nogalamycin, digitoxin, or antitumor agent (15). We also identified 2 terpenoid drugs, control for 30 days. We measured the size of tumors twice a pristimerin and triptonide, isolated from Chinese medici- week during this time (Fig. 3C). Both nogalamycin and nal plants, as potent inhibitors of growth (16–18). Among digitoxin showed more activity against Panc502 than the 125 antineoplastic drugs tested (Supplementary Table Panc410, supporting the notion that in vitro sensitivity does S2), only a third (49 drugs, 39.2%) showed activity above predict in vivo response, at least with these 2 drugs in these the bell-shaped curve. Nine drugs showed greater than 99% cell lines, as judged by tumor size (Fig. 3C), weight of the growth inhibition, including 7 nucleic acid–targeting che- tumors after completing the treatment (Fig. 3D), and by motherapeutics (actinomycin C, actinomycin D, chromo- visual inspection of the residual tumors after treatment mycin A3, camptothecin, doxorubicin, gemcitabine, and (Supplementary Fig. S7).

1142 Clin Cancer Res; 19(5) March 1, 2013 Clinical Cancer Research

Cancer Cell Line Isolating Tools

A B × 13.8 300 × 11.4 600 400 200 ND 400

100 (nmol/L) 200 Frequency 50 200 0 IC -25 025 50 75 100 0 0 +3SD 502 486 410 502 486 410 % growth inhibition Nogalamycin Digitoxin C D Panc410 Panc502 × 3.5 5 10 100 × 2.1 × 3.1 × 3.2 5 50 1 0 Tumor volume index 0 0 410 502 0 10 20 30 0 10 20 30 Normalized tumor weight Control Nogalamycin Digitoxin Days of treatment 0.2 mg/kg 1 mg/kg 0.4 mg/kg 2 mg/kg

Figure 3. Chemosensitivity of a low-passage familial pancreatic cancer from surgery. Histogram of the number of drugs (frequency) as a function of percentage of growth inhibition (A). Curve-fitting of Gaussian distribution onto the histogram (black line) distinguishes the distribution of drugs with little or no activity from those which show some level of activity above this distribution. Arrowhead indicates þ3SD above the mean. Arrows indicate the percentage of growth inhibition for nogalamycin (N) and digitoxin (D). Cell line–specific sensitivity of nogalamycin and digitoxin in the cell lines Panc502, Panc486, and Panc410 (B). Values shown are the mean IC50 values of 3 replicates and error bars are the 95% confidence intervals. In vivo growth curves of subcutaneous mouse xenografted tumors raised from the Panc410 and Panc502 cell lines after treatment with nogalamycin, digitoxin, or control (C). Control (empty ovals), nogalamycin 0.2 mg/kg (empty squares), nogalamycin 1.0 mg/kg (filled squares), digitoxin 0.4 mg/kg (empty triangles), and digitoxin 2.0 mg (filled triangles). Normalized weight of tumors explanted from mice after 30 days of treatment (D). Normalized tumor weight of Panc410 and Panc502 in white columns or gray columns, respectively. Error bars are SDs. Fold changes between Panc410 and Panc502 are noted.

Isolation of additional cell lines with a family history of pancreatic cancer (Supplementary To test whether we could use this system to routinely Fig. S9). In each case, cell lines were documented to be generate cell lines from solid cancers, we also isolated a cell tumorigenic in athymic mice, sequenced to document line from an ovarian cancer, another highly lethal cancer. oncogenic mutations in the Kras2 gene, and DNA ﬁnger- The ovarian cancer cell line, FM108, was established from printed to conﬁrm their patient origin (information avail- an existing xenograft (Supplementary Fig. S8). We also able on request). During this time, we isolated 3 additional isolated an additional cell line from a surgically resected cell lines from other PDAs but were unable to produce cell PDA. Panc486 was isolated from a xenograft from a patient lines from 2 other PDA xenografts using similar approaches

Table 1. Ten drugs with 75% or greater inhibition

Percentage of growth IC50, nmol/L inhibition at 10 mmol/L Drug on Panc502 Panc502 Panc486 Panc410 Triptolide (for triptonide) 98.1 7.3 10.2 8.3 Nogalamycin 75.4 23.3 265.2 281.9 Ouabain 95.6 39.7 57.9 81 Digitoxin 95.1 62.7 199.7 572.2 Strophanthin K 98.9 69.3 96.8 132.5 Digoxin 94.7 70.8 101.6 193.2 Dithiazanine iodide 99.2 234.1 156.4 283.3 Strophanthidin 98 430.1 549 988.9 Quinacrine 100 657.7 756.6 961.2 Celastrol (for pristimerin) 100 753.6 1229 1076

www.aacrjournals.org Clin Cancer Res; 19(5) March 1, 2013 1143

Kamiyama et al.

(total 7 cell lines, of 9 attempts, 78%). This suggests that the chosen, in part, because it only had 75.4% growth inhibi- in vitro propagation of the 2 failed cell lines is dependent on tion in vitro. Finally, it is possible that the drugs identified one of a number of factors such as factors intrinsic to the could be synergistic with those currently in clinical use, such carcinoma, factors related to the nonneoplastic stromal as gemcitabine, although this remains to be shown. In the cells, putative paracrine growth factors, or critical media future, the combination of the patient’s germline single- components that were missing and remain to be identified. nucleotide polymorphisms (SNP), somatic mutations, mRNA expression, and epigenetic changes in the cancer may play an increasing role in the choice of chemotherapy. Discussion Despite this, one can also imagine a role for functional We report successful production of nude, SCID, and NSG CSRAs. This is currently being attempted for various cancers immunodeficient mice whose hprt-null cells are biochem- (23–26), although it is controversial whether this testing is ically selectable in vitro. Using this system, we were able to truly predictive of in vivo response (2, 3). This initial sample isolate 3 PDA and 1 ovarian cancer cell line. The Panc502 took about 8 months from surgical resection to full drug familial PDA cell line was isolated directly from the tumor panel testing; however, we estimate that, with additional surgically resected from a patient, and we screened it against optimization, this testing could be accomplished much a panel of 3,131 drugs. Approximately 314 drugs show shorter time. The shorter time could be achieved by potential activity, where the vast majority of them (265, implanting more mice, with additional experience, and 84%) were not designated as antineoplastics. Of 125 anti- most importantly by using robotics for the drug screening. neoplastic drugs, only 49 (39%) showed activity. Two novel Certain clinical settings, such as identifying drugs after a drugs showed selective activity in vitro, and this correlated Whipple procedure for PDA, might be especially amenable with in vivo response. We also identified 2 terpenoid drugs to this time frame. Applying a similar approach to a cohort with more than 99% growth inhibition. of genetically defined cell lines could define a standardized In this article, we describe an alternative strategy to core set of drugs to be tested on all resected PDAs. traditional CSRAs. In contrast to traditional CSRAs, the Cell lines are useful for many other studies of cancer approach described herein uses: (i) pure low-passage biology as they: (i) can be expanded indefinitely, (ii) con- human cancer cells, (ii) cells that are viable and well- tain all of the gene mutations present in the patient’s adapted to in vitro growth, and (iii) larger drug panels. The primary cancer, (iii) can be manipulated in vitro by adding use of pure cancer cells means that any signal detected is due or eliminating genes, (iv) can be implanted in mice to test to the drug’s effect on the malignant cells rather than the effects of these manipulations on the ability to form stromal cell toxicity. Because the cancer cells are growing tumors (21, 27), and (v) carry few if any additional genetic and healthy, measured effects are a combined effect of the changes from the primary cancer (28). Cancer cell lines drug and the cancer cells and not due to synergistic effects of are also important for other in vitro functional studies, the drug with anoxia, starvation, overnight shipping, etc. molecular imaging, and possibly isogenic cancer vaccine Finally, because the cells are expanded into a low-passage production. With increasingly powerful tools such as cell line, drug panels with thousands of drugs can be whole-genome DNA sequencing, RNA interference librar- screened, rather than the 5 to 10 typically tested in tradi- ies, and high-throughput drug screens, one can anticipate tional CSRAs. This permits one to test all FDA-approved that the need for cell lines from human cancers will likely antineoplastics, FDA-approved drugs for other indications, increase (13, 29–31). late-phase clinical trial drugs, herbal drugs, and drugs We envision several applications of this system for cancer approved in other countries for clinical use. The idea of cell line production. First, there are no cell lines representing screening drugs using cancer cell lines is well established some malignancies, such as oligodendrogliomas and pheo- (reviewed in ref. 19), and the personalized approach has chromocytomas, and in others, such as esophageal, thyroid, recently been validated for one patient with a rare type of and salivary gland tumors, many cell lines are contaminants pancreatic cancer, acinar cell carcinoma (20). (32). For others, such as prostate, ovarian, and breast The approach describe herein differs from the xenograft- cancers, the number of available cell lines cannot fully based approach show by Hidalgo’s group (21, 22). The represent the full spectrum of disease (e.g., most existing advantage of the low-passage cell line approach is that more prostate cancer cell lines are hormone-refractory). In addi- drugs can be screened under highly controlled conditions. A tion, some cancers contain unique genetic defects, and possible advantage of the xenograft approach is that xeno- successful cell line production is essential to identify path- grafted tumor response may be more biologically predictive way members, and perhaps most importantly, for high- of response in patients. We are eager to test whether low- throughput drug screening. passage personalized cell line responses can predict patient There are some limitations of this system. For some response. This may be best tested with a malignancy cancers, stromal cells from the tumor microenvironment for which several alternative first-line chemotherapeutic may provide paracrine growth stimuli to the malignant choices are available. cells, and this may explain why we were unable to isolate Drugs with activity in vivo may not be limited to those that 2 of the cell lines. It is generally believed that all of the show more than 95% growth inhibition. While digitoxin human stromal cells are replaced by mouse stromal cells in a showed 95.1% growth inhibition in vitro, nogalamycin was single passage in immunodefective mice. While we

1144 Clin Cancer Res; 19(5) March 1, 2013 Clinical Cancer Research

Cancer Cell Line Isolating Tools

generally agree with this assessment based on phase micros- Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): H. Kamiyama, S. Rauenzahn, J.S. Shim, C.A. copy of HAT-treated cultures (unpublished data), this has Karikari, G. Feldmann, L. Hua, M. Kamiyama, F.W. Schuler, M.-T. Lin, R.M. never been formally quantified. The replacement of stromal Beaty, B. Karanam, H. Liang, G. Mo, H.A. Jinnah, A. Jimeno, J.R. Eshleman cells may be slightly less than 100% based on the extremely Analysis and interpretation of data (e.g., statistical analysis, biosta- tistics, computational analysis): H. Kamiyama, S. Rauenzahn, J.S. Shim, faint residual PCR product from DNA isolated from xeno- C.A. Karikari, M.-T. Lin, B. Karanam, R.H. Hruban, R.B.S. Roden, J.O. Liu, A. grafts of cancers containing homozygous deletions (33). In Maitra, J.R. Eshleman Writing, review, and/or revision of the manuscript: H. Kamiyama, S. this regard, it is possible that some tumors may need to be Rauenzahn, C.A. Karikari, M.-T. Lin, B. Karanam, M. Hidalgo, E. Jaffee, R.H. passaged through additional mice to get complete replace- Hruban, H.A. Jinnah, R.B.S. Roden, A. Jimeno, J.O. Liu, A. Maitra, J.R. ment of the human stromal cells. It is also possible that Eshleman Administrative, technical, or material support (i.e., reporting or orga- some cancers require unique cytokines that are not currently nizing data, constructing databases): S. Rauenzahn, J.S. Shim, G. Feld- components of tissue culture medium. mann, F.W. Schuler, H. Liang, M.E. Mullendore, H.A. Jinnah In addition to cell line production, one can envision Study supervision: J.R. Eshleman other situations where an animal model is required to study a biologic process but where recovery of the exogenous cells Acknowledgments is subsequently required. For example, the mouse may be The authors dedicate this work to the memory of Dr. Dawn Audi. They useful for isolation of organ-specific spontaneous metasta- thank Dr. Scott Kern for the suggestion to use hprt as the selectable marker (PCT/US01/31219), Dr. Ming-Sound Tsao for generously providing the ses or small numbers of dormant malignant cells refractory HPDE cell line, and Drs. Yoshihisa Matsushita, Bert Vogelstein, Cynthia to chemotherapeutic agents. Finally, high-throughput func- Zahnow, Julie Watson, Curt Civin, Cory Brayton, Don Price, Phil Wong, tional chemosensitivity profiling may be especially war- Craig Henke, Mehtab Khan, and Rajni Sharma for helpful discussions. ranted for those patients who have failed conventional chemotherapeutic options. Grant Support fl This work was funded in part from NIH grants CA130938 (J.R. Eshleman), Disclosure of Potential Con icts of Interest CA62924 (Drs. Scott Kern, R.H. Hruban, A. Maitra, J.R. Eshleman), S. Rauenzahn, A. Maitra, and J.R. Eshleman may receive royalty payments CA122581 (R.B.S. Roden), The Sol Goldman Pancreatic Cancer Research if the mice are licensed; a patent licensed to Myriad Genetics (R.H. Hruban Center, The Stewart Trust Fund, The Lustgarten Foundation, the Mary Lou and J.R. Eshleman); and Advisory Board membership in Roche Molecular Wootton Pancreatic Cancer Research Fund, The Michael Rolfe Pancreatic Diagnostics (J.R. Eshleman). No potential conflicts of interest were disclosed Cancer Foundation, and the HERA Foundation (R.B.S. Roden). by the other authors. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked Authors' Contributions advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate Conception and design: S. Rauenzahn, M. Hidalgo, R.H. Hruban, this fact. A. Maitra, J.R. Eshleman Development of methodology: S. Rauenzahn, J.S. Shim, C.A. Karikari, Received June 28, 2012; revised November 26, 2012; accepted December A. Maitra 14, 2012; published OnlineFirst January 22, 2013.

References 1. Wright JC, Cobb JP, Gumport SL, Golomb FM, Safadi D. Investigation 9. Hooper M, Hardy K, Handyside A, Hunter S, Monk M. HPRT-deficient of the relation between clinical and tissue-culture response to che- (Lesch-Nyhan) mouse embryos derived from germline colonization by motherapeutic agents on human cancer. N Engl J Med 1957;257: cultured cells. Nature 1987;326:292–5. 1207–11. 10. Jinnah HA, Page T, Friedmann T. Brain purines in a genetic mouse 2. Schrag D, Garewal HS, Burstein HJ, Samson DJ, Von Hoff DD, model of Lesch-Nyhan disease. J Neurochem 1993;60:2036–45. Somerfield MR. American Society of Clinical Oncology Technology 11. Littlefield JW. Selection of Hybrids from matings of fibroblasts in Vitro Assessment: chemotherapy sensitivity and resistance assays. J Clin and their presumed recombinants. Science 1964;145:709–10. Oncol 2004;22:3631–8. 12. Kohler G, Milstein C. Continuous cultures of fused cells secreting 3. Samson DJ, Seidenfeld J, Ziegler K, Aronson N. Chemotherapy sen- antibody of predefined specificity. Nature 1975;256:495–7. sitivity and resistance assays: a systematic review. J Clin Oncol 13. Chong CR, Xu J, Lu J, Bhat S, Sullivan DJ Jr, Liu JO. Inhibition of 2004;22:3618–30. angiogenesis by the antifungal drug itraconazole. ACS Chem Biol 4. Dangles-Marie V, Pocard M, Richon S, Weiswald LB, Assayag F, 2007;2:263–70. Saulnier P, et al. Establishment of human colon cancer cell lines from 14. Jones S, Zhang X, Parsons DW, Lin JC, Leary RJ, Angenendt P, et al. fresh tumors versus xenografts: comparison of success rate and cell Core signaling pathways in human pancreatic cancers revealed by line features. Cancer Res 2007;67:398–407. global genomic analyses. Science 2008;321:1801–6. 5. Gazdar AF, Kurvari V, Virmani A, Gollahon L, Sakaguchi M, Westerfield 15. Zhang H, Qian DZ, Tan YS, Lee K, Gao P, Ren YR, et al. Digoxin and M, et al. Characterization of paired tumor and non-tumor cell lines other cardiac glycosides inhibit HIF-1alpha synthesis and block tumor established from patients with breast cancer. Int J Cancer 1998;78: growth. Proc Natl Acad Sci U S A 2008;105:19579–86. 766–74. 16. Chang FR, Hayashi K, Chen IH, Liaw CC, Bastow KF, Nakanishi Y, et al. 6. Jaffee EM, Schutte M, Gossett J, Morsberger LA, Adler AJ, Thomas M, Antitumor agents. 228. five new agarofurans, Reissantins A-E, and et al. Development and characterization of a cytokine-secreting pan- cytotoxic principles from Reissantia buchananii. J Nat Prod 2003;66: creatic adenocarcinoma vaccine from primary tumors for use in clinical 1416–20. trials. Cancer J Sci Am 1998;4:194–203. 17. Phillips PA, Dudeja V, McCarroll JA, Borja-Cacho D, Dawra RK, 7. Phelps RM, Johnson BE, Ihde DC, Gazdar AF, Carbone DP, McClin- GrizzleWE,etal.Triptolideinduces pancreatic cancer cell death tock PR, et al. NCI-Navy Medical Oncology Branch cell line data base. via inhibition of heat shock protein 70. Cancer Res 2007;67: J Cell Biochem Suppl 1996;24:32–91. 9407–16. 8. McBain JA, Weese JL, Meisner LF, Wolberg WH, Willson JK. Estab- 18. Titov DV, Gilman B, He QL, Bhat S, Low WK, Dang Y, et al. XPB, a lishment and characterization of human colorectal cancer cell lines. subunit of TFIIH, is a target of the natural product triptolide. Nat Chem Cancer Res 1984;44:5813–21. Biol 2011;7:182–8.

www.aacrjournals.org Clin Cancer Res; 19(5) March 1, 2013 1145

Kamiyama et al.

19. Sharma SV, Haber DA, Settleman J. Cell line-based platforms to 26. Ugurel S, Schadendorf D, Pfohler C, Neuber K, Thoelke A, Ulrich J, evaluate the therapeutic efﬁcacy of candidate anticancer agents. Nat et al. In vitro drug sensitivity predicts response and survival after Rev Cancer 2010;10:241–53. individualized sensitivity-directed chemotherapy in metastatic mela- 20. Armstrong MD, Von Hoff D, Barber B, Marlow LA, von Roemeling C, noma: a multicenter phase II trial of the Dermatologic Cooperative Cooper SJ, et al. An effective personalized approach to a rare tumor: Oncology Group. Clin Cancer Res 2006;12:5454–63. prolonged survival in metastatic pancreatic acinar cell carcinoma 27. Jimeno A, Amador ML, Kulesza P, Wang X, Rubio-Viqueira B, Zhang X, based on genetic analysis and cell line development. J Cancer et al. Assessment of celecoxib pharmacodynamics in pancreatic 2011;2:142–52. cancer. Mol Cancer Ther 2006;5:3240–7. 21. Rubio-Viqueira B, Jimeno A, Cusatis G, Zhang X, Iacobuzio-Donahue 28. Jones S, Chen WD, Parmigiani G, Diehl F, Beerenwinkel N, Antal T, C, Karikari C, et al. An in vivo platform for translational drug develop- et al. Comparative lesion sequencing provides insights into tumor ment in pancreatic cancer. Clin Cancer Res 2006;12:4652–61. evolution. Proc Natl Acad Sci U S A 2008;105:4283–8. 22. Hidalgo M, Bruckheimer E, Rajeshkumar NV, Garrido-Laguna I, De 29. Wood LD, Parsons DW, Jones S, Lin J, Sjoblom T, Leary RJ, et al. The Oliveira E, Rubio-Viqueira B, et al. A pilot clinical study of treatment genomic landscapes of human breast and colorectal cancers. Science guided by personalized tumorgrafts in patients with advanced cancer. 2007;318:1108–13. Mol Cancer Ther 2011;10:1311–6. 30. Hannon GJ. RNA interference. Nature 2002;418:244–51. 23. Eltabbakh GH, Piver MS, Hempling RE, Recio FO, Lele SB, Marchetti 31. Hudson TJ, Anderson W, Artez A, Barker AD, Bell C, Bernabe RR, et al. DL, et al. Correlation between extreme drug resistance assay and International network of cancer genome projects. Nature 2010;464: response to primary paclitaxel and cisplatin in patients with epithelial 993–8. ovarian cancer. Gynecol Oncol 1998;70:392–7. 32. Boonstra JJ, van Marion R, Beer DG, Lin L, Chaves P, Ribeiro C, et al. 24. Fruehauf JP, Alberts DS. Assay-assisted treatment selection for wom- Veriﬁcation and unmasking of widely used human esophageal ade- en with breast or ovarian cancer. Recent Results Cancer Res 2003; nocarcinoma cell lines. J Natl Cancer Inst 2010;102:271–4. 161:126–45. 33. Hahn SA, Seymour AB, Hoque AT, Schutte M, da Costa LT, Redston 25. Furukawa T, Kubota T, Hoffman RM. Clinical applications of the MS, et al. Allelotype of pancreatic adenocarcinoma using xenograft histoculture drug response assay. Clin Cancer Res 1995;1:305–11. enrichment. Cancer Res 1995;55:4670–5.

1146 Clin Cancer Res; 19(5) March 1, 2013 Clinical Cancer Research

Personalized Chemotherapy Profiling Using Cancer Cell Lines from Selectable Mice

Hirohiko Kamiyama, Sherri Rauenzahn, Joong Sup Shim, et al.

Clin Cancer Res 2013;19:1139-1146. Published OnlineFirst January 22, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-12-2127

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2013/01/22/1078-0432.CCR-12-2127.DC1

Cited articles This article cites 33 articles, 16 of which you can access for free at: http://clincancerres.aacrjournals.org/content/19/5/1139.full#ref-list-1

Citing articles This article has been cited by 5 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/19/5/1139.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/19/5/1139. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.